Bowen Baker

And

We hope that if we make a broad enough suite of evaluations that test monitorability in different ways, that when a model does start to become obfuscated or verbalize less for whatever reason, we'll see one of those evals start to go down.

And we did kind of show a proof of concept of that in the paper where we directly, deliberately trained models to obfuscate by putting some kind of style-like pressures on the COT.

And we then did see...

some of our evals decrease as like a proof of concept that, you know, if a model, if this kind of happened in the wild naturally, we'd maybe be able to catch it.

Totally possible.

I guess if I knew what the next scaling paradigm was, I'd be a much richer man.

Yeah, I think that is definitely a worry.

We've also like, I mean, so generally in our current experiments, we haven't necessarily found that to be true to a significant extent.

I think that this is likely because, you know, we are always improving our ability to train even, you know, even a small model.

It's not like we have our GPT-4 size model and that's the best GPT-4 size model anyone could ever train.

We're like improving our algorithms, our data.

you know, no one's sleeping.

And so I think that, you know, a better model also kind of like has more clear thoughts sometimes, if that makes sense.

Like a small model, if you look at its thoughts, it sometimes feels a lot more gibberish-y than a bigger model.

And so, you know, that is maybe the reason why we could expect this to, you know, at least

not fully, this trend to not fully happen.

But I think at the limit, it probably should happen.

Like bigger models will just verbalize less for the most part.

Yeah.